[clang] [llvm] [RISCV][P-ext] Support Packed "Q-format" Multiply Parts Accumulate (PR #217918)

[via cfe-commits]

================
@@ -12347,12 +12422,24 @@ SDValue RISCVTargetLowering::LowerINTRINSIC_WO_CHAIN(SDValue Op,
     switch (IntNo) {
     case Intrinsic::riscv_orc_b:      Opc = RISCVISD::ORC_B;      break;
     case Intrinsic::riscv_brev8:      Opc = RISCVISD::BREV8;      break;
-    case Intrinsic::riscv_sha256sig0: Opc = RISCVISD::SHA256SIG0; break;
-    case Intrinsic::riscv_sha256sig1: Opc = RISCVISD::SHA256SIG1; break;
-    case Intrinsic::riscv_sha256sum0: Opc = RISCVISD::SHA256SUM0; break;
-    case Intrinsic::riscv_sha256sum1: Opc = RISCVISD::SHA256SUM1; break;
-    case Intrinsic::riscv_sm3p0:      Opc = RISCVISD::SM3P0;      break;
-    case Intrinsic::riscv_sm3p1:      Opc = RISCVISD::SM3P1;      break;
+    case Intrinsic::riscv_sha256sig0:
----------------
TelGome wrote:

 Sorry, that was an unrelated formatting change. Reverted.

https://github.com/llvm/llvm-project/pull/217918